Examining the Clinical Prediction of Cohesion in Group Psychotherapy and Potential Moderators

Group cohesion, or the way in which group members relate and work toward the primary task of a therapy group, is a predictor for positive outcome in group psychotherapy and the building block for most group development models. Research has shown that interventions are most effective when tailored for the developmental stage of the group (Burlingame, Fuhriman, & Johnson, 2001). Logic follows that if best practice dictates that interventions be implemented based on developmental stage which, in turn, is largely informed by group cohesion, then group leaders should be competent in accurately assessing the cohesion of the group. To date, no study has examined the relationship between group leaders’ perceptions of cohesion compared to group members’ perceptions of cohesion. Further, there is little research on member or leader variables that may lead to more (or less) congruence between group leader and group member reports of cohesion. This study utilized hierarchical linear modeling (HLM) to examine the relationship between leader and member scores on a measure of group cohesion. Several potential moderators were also included in the analysis to test for interaction effects between group leader and group member scores of cohesion. Moderators of interest included group member vulnerability, group leader experience, the amount of “here-and-now” processing done in group sessions (as reported by the leader), and the number of completed sessions at the time of data collection. The study analyzed 103 total group members nested within 21 preexisting psychotherapy groups from community mental health centers, college counseling centers, university training clinics, and private practices in the Rocky Mountain region. Results of the study showed that group leaders across groups consistently reported a lower level of cohesion compared to group member reports. Due to this incongruence between leaders and members, group discussion of cohesion or use of group cohesion measures could aid group members and the group leader in understanding their group’s dynamics similarly. While none of the moderators examined reached statistical significance in the moderation effect of group leader and group member cohesion scores, the amount of “here-and-now” processing done in group as reported by the group leader showed most promise as a possible moderator. Future research examining additional variables that may predict greater congruence between leader and member reports of cohesion can further inform both research and practice in group psychotherapy